INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PITUITARY GROWTH HORMONE (GH) secretion is stimulated by hypothalamic growth hormone-releasing hormone (GHRH) and is inhibited by somatostatin. The release of GH is pulsatile with a 3- to 4-h interval between GH surges in the male rat (46). The pulses correlate with the occurrences of nonrapid eye-movement sleep (NREMS) in the rat (27), whereas the major daily GH release is coupled to deep NREMS in humans (reviewed in Ref. 50).

GHRH also promotes sleep. The somnogenic activity of GHRH has been verified by intracerebroventricular injections in rats and rabbits (13, 31, 32) and by systemic administration in humans (21, 24, 43) and rats (33). GHRH consistently stimulates NREMS; the promotion of rapid eye-movement sleep is variable. GHRH also enhances slow-wave activity in the electroencephalogram (EEG) during NREMS (13, 31, 32). Inhibition of endogenous GHRH by means of a competitive antagonist (34) or immunoneutralization (35) decreases sleep. NREMS also decreases in transgenic mice deficient in the somatotropic axis including GHRH (54). NREMS is rapidly suppressed in response to somatostatinergic stimulation in the rat (2). Hypophysectomy fails to inhibit the NREMS-promoting activity of GHRH (33). It has been suggested that stimulation of GH secretion and NREMS are independent though normally synchronized functions of GHRH. Promotion of NREMS is attributed to GHRHergic neurons projecting to the basal forebrain, where GHRH acts as a neurotransmitter-neuromodulator (11, 25, 41).

Long periods of wakefulness elicit sleepiness followed by enhanced sleep. Sleep deprivation (SD) is therefore a widely used tool to stimulate the sleep process. The duration, continuity, and intensity of NREMS all increase during recovery (8, 15, 26, 36). Enhancements in NREMS precede increases in rapid eye-movement sleep after SD (16, 47). The intensity (depth) of NREMS is characterized by the slow-wave activity in the EEG (7). In the normal, undisturbed rhythm of sleep-wake activity, high-intensity NREMS occurs during the first portion of the sleep period immediately following the circadian active phase (5, 8, 22). Indeed, there is a diurnal rhythm in hypothalamic GHRH mRNA levels (10) and SD-induced changes in hypothalamic GHRH mRNA also occur in the rat (48, 53). However, hypothalamic GHRH and somatostatin contents at various time points of the diurnal cycle and during and after SD have not heretofore been determined. The results indicate that changes in GHRH but not somatostatin correlate with NREMS.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Adult male Sprague-Dawley rats (body weights 300-320 g) were acclimated to a 12:12-h light-dark cycle (light on 0800) by housing them in individual cages placed in environmental chambers for at least 2 wk before the experiments. Ambient temperature was regulated at 26 ± 1°C. To study the diurnal rhythm in hypothalamic peptide content the rats were killed at 4-h intervals starting 1 h after light onset. In the experiments with SD, the rats were sleep deprived by gentle handling starting at light onset. Rats were killed at the end of 4 (n = 9) and 8 h (n = 9) of SD and after 1 (n = 10) and 2 h (n = 9) of recovery following the termination of SD. The time-matched control rats were undisturbed and killed at hours 4, 7, 9, and 10 of the light period. In addition, one group of rats was killed at light onset (hour 0). To help in the determination of the diurnal rhythms of the peptides, the results obtained from the controls of the SD groups are plotted together with the data collected in the diurnal rhythm experiments so that the peptide contents are depicted for hours 0 (n = 10), 1 (n = 10), 4 (n = 8), 5 (n = 11), 8 (n = 9), 9 (n = 19), and 10 (n = 10) in the light period and hours 13 (n = 12), 17 (n = 11), and 21 (n = 10) in the dark period. In both experiments, only 1 to 3 rats were killed simultaneously at a particular time point in one session, and the sessions were repeated until a sample size consisting of a minimum of eight rats was obtained for each time point. A guillotine was used to kill the rats. The hypothalamus (landmarks: optic chiasma, lateral sulci, mammillary bodies, and a depth of 2 mm) was removed within 1 min and stored at 70°C until assayed. The landmarks for the dissection of the hypothalamus corresponded to those used by Katakami et al. (19) and Gil et al. (18) to measure hypothalamic GHRH contents. Calculated with respect to hypothalamic samples, our GHRH values (1,800-2,800 pg/hypothalamus) are in the range with those reported by Gil et al. (18), approximately 1,800-1,900 pg/hypothalamus, but higher than the hypothalamic GHRH content, 900 pg/hypothalamus, found by Katakami et al. (19). The differences can be due to differences in the antisera and the reference peptide.

The frozen samples were weighed and placed in tubes containing 0.5 ml of 2 M acetic acid and then boiled for 5 min (19). The tissues were individually homogenized by means of ultrasound. The homogenates were centrifuged at 15,000 g for 20 min at 4°C. Aliquots were taken for protein measurements (9), and the rest of each supernatant was lyophilized. The recoveries of the extraction were 70-75% for GHRH and 90-95% for somatostatin. Radioimmunoassay was used to determine GHRH and somatostatin contents. The kits were purchased from Peninsula Laboratories (standards rat GHRH 43 and somatostatin 1-14). According to the supplier, the antiserum to rat GHRH exhibited 100% cross-reactivity with human GHRH, whereas it did not recognize peptides as follows: His¹-Nle²⁷-GHRH 132, porcine GHRH, human parathyroid hormone 134, rat peptide histidine-isoleucine, and human, rat, and porcine vasoactive intestinal peptide. There was a 100% cross-reaction between the antiserum to somatostatin and somatostatin 25, somatostatin 28, and [Des-Ala¹]-somatostatin, 0.05% cross-reaction with [D-Trp⁸]-somatostatin, and 0.01% cross-reaction with prosomatostatin 32. The antiserum to somatostatin failed to recognize the somatostatin analogs RC-160 and CTOP-NH₂, human amylin amide, human glucagon, human insulin, porcine neuropeptide Y, substance P, and human, rat, and porcine vasoactive intestinal peptide. The peptides were measured in triplicate with a sensitivity of 8 pg/tube for GHRH and 4 pg/tube for somatostatin. The intra- and interassay coefficents of variation were 4.5 and 5.0 for GHRH and 3.4 and 11.7 for somatostatin, respectively.

The peptide content of the individual samples were averaged for each time point. One-way analysis of variance (ANOVA) was used to detect significant differences among the various time points across the 24-h cycle. The changes in GHRH and somatostatin during and after SD were evaluated with respect to the time-matched control samples by means of two-way ANOVA with the treatment (SD vs. controls) and the time (hours after light onset) as the two factors. Student's t-tests were used to compare mean peptide contents between the light and dark periods, and Pearson Product Moment Correlation was calculated to compare variations in tissue mass and GHRH somatostatin contents. An alpha-level of P < 0.05 was considered to be significant in all tests.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The wet weights of the hypothalamic samples varied between 25 and 30 mg without significant differences among the individual time points. In contrast, significant variations were observed in the protein contents among the samples [F(9,100) = 3.42; P < 0.05]. This finding is consistent with previous reports on diurnal variations in hypothalamic protein contents (1, 28). Although the patterns of variations in GHRH and somatostatin contents expressed as picogram per milligram tissue or nanogram per microgram protein did not differ, only peptide contents expressed as picogram per milligram tissue (wet weight) are presented and will be discussed.

GHRH. Hypothalamic GHRH content exhibited significant diurnal variations [F(9,100) = 3.29; P < 0.05] (Fig. 1). GHRH content was low at light onset. A transient rise in GHRH was observed in hour 1 after light onset, and then GHRH dropped to a minimum occurring by the middle of the day. GHRH content increased gradually during the second portion of the light period. Peak values were observed at the end of the light period and at the beginning of the dark period. GHRH declined steadily during the night. A transient rise in GHRH content was observed in hour 1 after light onset, which was followed by another decline. Minimum GHRH content occurred in the middle of the day. The mean GHRH values during the 12-h light and 12-h dark periods did not differ [light 74.8 ± 3.19 (SE) pg/mg, dark 81.5 ± 3.2 pg/mg, Student's t-test]. There were no correlations between the wet mass of the hypothalamus and the GHRH content in either the diurnal rhythm or SD experiments.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Diurnal ( and lines) and sleep deprivation (SD)-induced (bars) variations in hypothalamic growth hormone-releasing hormone (GHRH) content. Means ± SE are depicted for sample sizes between a minimum of 8 and a maximum of 19 at each time point. Diurnal rhythm is double plotted, and effects of SD are shown on day 2. Periods between hour 0 and hour 12, as well as hour 24 to hour 36 are light periods. Dark periods are marked by horizontal bars. SD started at light onset (hour 24) and lasted for 8 h (hour 32), as indicated by arrows. GHRH was determined after 4 and 8 h of SD, and after 1 and 2 h of recovery (R) after deprivation.

Somatostatin. Somatostatin content of the hypothalamus also varied significantly across the 24-h day [F(9,99) = 4.53; P < 0.05] (Fig. 2). Starting from a low value at light onset, a transient rise occurred in somatostatin content 1 h after light onset and then somatostatin declined. These variations were similar to those observed for GHRH. Somatostatin contents stayed low in the afternoon and started to rise only around dark onset. The peak of somatostatin content occurred 1 h after dark onset. Somatostatin declined rapidly at night and dropped to low levels by morning. Mean somatostatin contents during the light period [1,576.0 ± 60.6 (SE) pg/mg] were significantly lower than at night (2,031.6 ± 129.0 pg/mg). Somatostatin did not correlate with hypothalamic mass. Changes in somatostatin failed to correlate with the variations in hypothalamic GHRH content (Pearson Product Moment Correlation 0.6179; P = 0.0569).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Diurnal ( and lines) and SD-induced (bars) variations in hypothalamic somatostatin contents. See Fig. 1 for additional data.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Current results demonstrate diurnal variations and SD-induced changes in hypothalamic GHRH content. In rats the duration and the intensity of NREMS peak at the beginning of the light period. Both of these parameters of NREMS decline toward dark onset. Rats sleep relatively little at night, but the intensity of NREMS increases (5, 8, 22). The SD protocol used in our experiments induces recovery sleep with intense NREMS for a few hours (47). Therefore, the periods that are likely to be associated with simultaneous increases in GHRH release and synthesis, the first portion of the light period and recovery, correlate with the period of most intense NREMS. It is assumed that the intensity of NREMS is a function of prior wakefulness (7). Enhancements in EEG power densities (15) and increases in EEG amplitudes (16) can already be demonstrated in the waking EEG during SD and are regarded as markers of the homeostatic process resulting in the sleep drive. If depletion of hypothalamic GHRH indicates GHRH release at night and during SD then GHRH may be part of the homeostatic process responsible for somnolence. It is noted that the SD-induced variations in GHRH might be due in part to stress associated with the procedure of SD. If the stress component exists, then it is inherently linked to SD and it is also involved in the mechanism of sleep alterations. Irrespective of whether the changes in GHRH release result from the sleep loss per se or from stress, the role of GHRH in the recovery sleep after deprivation is supported by the previous observation that immunoneutralization of GHRH blocks or attenuates the recovery sleep (35).

Promotion of sleep is attributed to GHRHergic neurons residing outside of the arcuate nucleus. These neurons project to various structures in the basal forebrain, including the medial preoptic area (25, 41). Several lines of evidence indicate that this region has a fundamental role in the regulation of NREMS (reviewed by Ref. 44). The basal forebrain is also implicated in the mediation of the effects of GHRH on NREMS, e.g., microinjection of GHRH into the medial preoptic area enhances sleep (55). Studies using in situ hybridization (48) demonstrate that both the diurnal variations and the SD-induced changes in GHRH mRNA occur in extra-arcuate GHRHergic neurons. There is, however, an interesting dissociation between these events: diurnal variations are observed in the periventromedial GHRHergic neurons, whereas SD increases GHRH mRNA in the paraventricular nucleus. It is possible therefore that different GHRHergic neurons provide circadian and homeostatic inputs for sleep regulation.

Considering the opposite effects of GHRH and somatostatin on both GH secretion and sleep, it was an attractive idea to determine both peptides from the same hypothalamic samples. Unlike GHRH, however, somatostatin is a pleiotropic neurotransmitter in interneurons involved in many functions throughout the hypothalamus (reviewed in Ref. 6). Somatostatin content was not significantly altered by SD, although somatostatin displayed diurnal variations. Previously, Nicholson et al. (30) reported that hypothalamic somatostatin content declines during the first portion of the light period. Berelowitz et al. (3) described progressive in vitro somatostatin release during the light period in hypothalamic samples obtained at various time points. Somatostatin concentrations in the cerebrospinal fluid and in hypothalamic portal vessels are higher at night than during the day (4); the large drop in somatostatin content of the hypothalamus might reflect this enhanced release. The diurnal rhythm of somatostatin content in our hypothalamic samples differ from the variations in somatostatin in the suprachiasmatic nucleus (42) or the anterior hypothalamus, which includes the periventricular area, e.g., the major source of the hypophyseotropic somatostatinergic neurons (17). Despite the obvious similarities in their diurnal oscillations, statistically significant correlations could not be demonstrated between the fluctuations of GHRH and somatostatin. Current findings do not exclude the possibility that somatostatin also contributes to sleep regulation. Somatostatinergic stimulation alters sleep (2), and we detected reciprocal changes in the somatostatin and GHRH mRNA levels in the hypothalamus in response to SD (53). The interaction between somatostatin and GHRH is also well documented in the regulation of GH secretion. Our data, however, show that hypothalamic somatostatin contents cannot be simply linked to the activity of the somatotropic axis or sleep without reference to specific nuclei.

Although the kinetics of translation, and those of peptide release, degradation and elimination are not known, comparisons between GHRH mRNA levels and peptide contents may allow some reasonable conclusions. Hypothalamic GHRH mRNA peaks around light onset, declines gradually throughout the light period, and stays at very low levels at night (10). Therefore, the transient rise in GHRH content in the morning may indicate a period of peptide synthesis. In our previous mRNA studies, samples were taken 3 h before light onset and 1 h after light onset. GHRH mRNA peaked 1 h after light onset. An instantaneous translation is unlikely. It is assumed therefore that peak transcription occurs 1 to 2 h earlier than the transient rise in peptide content. The finding that hypothalamic GHRH mRNA is already enhanced at light onset (48) supports this interpretation. A 1- to 2-h interval between the increases in GHRH mRNA and peptide contents corresponds to the kinetics of translation of GHRH in the arcuate nucleus (52). The drop in GHRH content after the morning peak indicates a rapid release exceeding the rate of synthesis until mid-day. The late afternoon rise in GHRH content is attributed to a combination of a continual peptide synthesis and a low rate of release. The decline in GHRH content at night may result from a release or degradation without significant synthesis as indicated by the steady, low level of GHRH mRNA.

Increases in hypothalamic GHRH mRNA are detected after SD starting at light onset and lasting for 6 (48), 8, or 12 h (53). GHRH content decreases significantly by the end of an 8-h SD, and stays low for 1 to 2 h during recovery. Considering that GHRH mRNA is already higher than normal at least 2 h before the termination of the 8-h SD then the low postdeprivation peptide content may occur despite a high rate of synthesis. Unfortunately, it is not known whether translation is altered during SD. Stimulation of transcription during SD may be a regulatory response to the depletion of GHRH sometime between hour 4 when the GHRH content is normal and hour 6 when GHRH mRNA is already enhanced. It is noted that the SD-induced changes in GHRH mRNA may differ between the light and dark periods. Thus Toppila et al. (48) report that a 12-h SD during the dark period fails to stimulate GHRH mRNA. GHRH mRNA, however, increases by light onset due to the diurnal rhythm, and this may mask the effects of SD.

Secretion of GH is pulsatile with surges occurring at 3- to 4-h intervals throughout the day in male rats (46). GH surges are timed to releases of GHRH into the pituitary portal vessels (37). In rats, light has a synchronizing effect on the ultradian rhythm of GH secretion, but there are no diurnal variations in GH release (46, 51). GH secretion tends to be suppressed during SD in humans (40, 45) and rats (20), although there are some inconsistencies among the reports (39), and the deprivation-associated decreases in plasma GH concentrations might depend on age (29). There is therefore no evidence implicating the observed diurnal variations or SD-associated changes in hypothalamic GHRH contents or GHRH mRNA in the regulation of GH secretion.

That the diurnal or SD-induced changes in hypothalamic GHRH contents and GHRH mRNA are not linked to GH secretion is supported by fundamental observations in an in situ hybridization experiment by Toppila et al. (48). The hypophyseotropic GHRH neurons reside in the arcuate nucleus (11, 41). The arcuate nucleus is also the structure where 3-h oscillations are observed in GHRH mRNA in correlation with GH release (52). Neither the light-dark cycle nor SD alters GHRH mRNA in the arcuate nucleus (48). In contrast, SD results in increases in somatostatin mRNA levels in the arcuate nucleus (48). The somatostatinergic cells in the arcuate nucleus are intranuclear interneurons (6) that inhibit GHRH release (23). It seems therefore that the activity of the hypophyseotropic GHRH neurons is suppressed rather than enhanced during SD.

In addition to the role of hypothalamic GHRH in the regulation of GH secretion and NREMS, GHRH is also implicated in the regulation of food intake (49). Food intake shows diurnal variations in the rat with most feeding occurring at night and a sharp drop in feeding activity around light onset followed by a gradual increase during the rest of the day (38). There are correlations between meal size and intermeal sleep (12). Food intake increases during chronic SD lasting for weeks (14). We are, however, not aware of any reports describing feeding during short-term SD as that used in our experiments, although postdeprivation food intake is not altered by short-term SD (8). Bouts of feeding occurred during SD in our rats, but qualitatively this activity did not seem to be different from feeding occurring periodically during the day. It therefore remains to be determined if variations in hypothalamic GHRH contents correlate with feeding activity.

In conclusion, the diurnal and SD-induced variations in hypothalamic GHRH content correlate with NREMS. Previous reports on diurnal and SD-associated variations in hypothalamic GHRH mRNA promote understanding of the changes in peptide content in terms of synthesis and release. The findings are consistent with the suggested role of GHRH in the promotion of NREMS.